Dr. Oehler's objective over 5 years is to prospectively assess mechanisms of resistance to imatinib mesylate in chronic phase (CP) CML patients and investigate the biology of relapse in patients who have obtained at least a MCR. Imatinib has become first line therapy based on the IRIS study results. However, the durability of response is unknown, and allogeneic transplant remains the only potential cure. Given the rapid change in treatment strategy for CML there is a clear need to identify predictors of response early in therapy, ideally pre-treatment. Specific Aim 1. Assess molecular response and clinical predictors of response to imatinib. The kinetics of bcr-abl disappearance as assessed by RT-PCR predict relapse and response to allogeneic transplant and interferon therapy. The Hasford score has been utilized to assess prognosis in patients treated with interferon or chemotherapy. All patients assigned to the four arms of SWOG S0325 will be assessed by RTPCR. We hypothesize that the kinetics of bcr-abl decrease within three to six months of initiating treatment will correlate with duration of response, and that the Hasford score predicts early response to imatinib. Specific Aim 2. Determine the genetic predictors of response and relapse by genomic expression analysis. Microarray technology has been used to develop genomic profiles that may be more predictive of prognosis than pathology or cytogenetics. Pre-treatment genomic predictors of response will be identified in chronic phase CML patients by microarray analysis. These predictors will be validated prospectively on SWOG S0325. We hypothesize that imatinib responders will have different expression profiles than non-responders and that low, intermediate, and high risk Hasford scores will have distinct gene expression profiles that correlate with outcome. Specific Aim 3. Determine the biology of relapse by point mutation analysis and genomic expression analysis. Several mechanisms of resistance to imatinib have been described. The prevalence of point mutations will be assessed by PCR-SSCP and by a sensitive mismatch assay to determine the frequency of point mutations pre- and early in treatment. We hypothesize that mutations frequency is high and is detectable before treatment. The gene expression pattern of the relapsed sample will be compared with the corresponding diagnostic sample. We hypothesize that the evolution of a previously undetected clone is the mechanism of relapse.